Paper for pre-delivery distribution, Fairbank Center for Chinese Studies, Harvard University (October 28, 1996). Not to be cited without author's permission.

For encapsulating a worldview there is nothing quite like a world map. As with other forms of cartography, mappaemundi--whether medieval or modern, Asian or Western--tell us about values and attitudes, aims and aspirations, hopes and fears; but they express them on a particularly grand, indeed global, scale. To the extent that such productions in any given society share affinities across space and time, they reveal significant features of that culture's self-image (and, of course, its conceptions of the "other"); and to the degree that they do not, they suggest changes, ruptures, tensions, and conflicts within the larger cultural system. With these considerations in mind, I would like to look at the evolution of Chinese maps of the world during late imperial times--from the twelfth to the twentieth centuries--focusing on two basic questions: How did changing conceptions of "the world" shape the contours of Chinese cartography, and how did changing (as well as enduring) cartographic practices affect Chinese conceptions of the world?

Significant methodological and practical problems attend such questions. In the first place, it is often difficult to determine where a map of "China" ends and a map of "the world" begins. Large-scale cartographic representations of space in late imperial times present us with a number of overlapping political, cultural, and geographical images, identified either by dynastic names (Song, Yuan, Ming, and Qing) or by more enduring, but still fluid, designations such as the Central [Cultural] Florescence (Zhonghua), the Spiritual Region (Shenzhou), the Nine Regions (Jiuzhou), the Central Kingdom (Zhongguo), the Central Land (Zhongtu), and All under Heaven (Tianxia). The relationship--as well as the distinction--between these time-honored concepts is by no means always clear in traditional Chinese maps.

Another difficulty has to do with access. Not surprisingly, a number of politically sensitive maps that deal with disputed territory are still not generally available to foreign scholars doing cartographic research in the People's Republic of China. Moreover, as far as I know, there is no single, comprehensive collection of Chinese "world maps" anywhere. Rather, they are scattered all over the globe--not only in Asian archives (primarily China, Taiwan, Japan, and Korea), but also in Russia, various European countries (including Italy, Spain, France, Germany, England, and Holland), and North America.

In the course of my research, I have been able to examine well over a hundred Chinese manuscript maps and woodblock prints in various Asian, European and American collections. I have also consulted a number of other valuable primary materials in Chinese, dating from the Song through the Qing dynasties, and have investigated much of the large and growing body of secondary literature on Chinese cartography in both Asian and Western languages. Still, this chapter must be seen as only a preliminary investigation, an initial voyage of exploration into a vast and potentially very productive field of study.

A number of scholars, both within and outside the field of Chinese studies, have recently criticized what is generally described as a Parsonian version of culture (i.e., culture as "systems of symbols and meanings") for contributing to various "totalizing" and "essentializing" orientalist projects, including the rise of "academic modernization theory" and "imperialist development policy." It has been blamed for creating a "neat divide between 'Oriental' culture and 'Western' reason," and for providing "the most convenient" explanation for the "willful backwardness and irrationality [of so-called traditional societies] in the face of rapid global modernization." In the view of critics such as Judith Farquhar and James Hevia, the reification of ideas and values encouraged by Parsons and his disciples has led to a "static and stagnant" conception of culture which justifies Western aggression and represents imperialism as "a salvation project"

It is not clear to me that all, or even most, of those who have employed some sort of Parsonian notion of culture in their academic writing (including John Fairbank within the China field and Clifford Geertz without) are guilty of such crimes. Nor am I convinced that the long-posited relationship between ideas, values, intentions, ideologies and other forms of consciousness on the one hand and human behavior or "action" on the other is wrong-headed. One can argue, I think, for positioning culture "in the materiality and (messiness) of everyday life"without disengaging it entirely from the realm of thought.

My own work on Chinese culture has tended more toward the generalizing than the particularizing side of the interpretive spectrum. I would hope, however, that it is not viewed as either a "totalizing" or an "imperialist" enterprise, and certainly not one that valorizes or privileges Western culture in any way. As I have tried to indicate in the most recent edition of China's Cultural Heritage (1994), my interest remains in sustaining a dialectic between holistic and particularistic studies of China. Pamela Crossley is assuredly correct in maintaining that the geographical and cultural entity of China is "a totality of convergently and divergently related localisms," and that Chinese culture is a product of the "challenging and differentiating effects effects of aboriginal, border and heterodox cultures" But simply to speak of aboriginal, border and heterodox cultures is to acknowledge implicitly a hegemonic, "central," and "orthodox" one in constant tension with them. What, one may legitimately ask, is the nature of that larger culture?

My interest is in the idea of culture as "classification"--that is, the way groups of Chinese (whether they see themselves primarily as "the Han people," "Guangdong people," or whatever) name and arrange things and ideas into coherent systems of meaning. In this respect I identify with the interpretive outlook of Marshall Sahlins. I am perfectly willing to acknowledge, however, that the "Chinese" cultures under consideration are neither static nor monolithic. They vary across time and space and according to class and gender; they are constantly changing, and are always situated in particular social and political contexts. I would also grant that these cultures are invariably the product of some sort of "invention," and the cultural meanings produced are constantly contested by different groups and individuals.

Yet there remains a sense in which people share an identity that can not only be encapsulated by one or more self-referential terms (as opposed to designations imposed from without) but also described (again from within) as a constellation of commonly accepted attributes, attitudes, and concerns. As Sahlins points out, "In order for categories to be contested . . . there must be a common system of intelligibility, extending to the grounds, means, modes, and issues of disagreement." It would be difficult, he argues, "to understand how a society could function, let alone how any knowledge of it could be constituted, if there were not some meaningful order in the differences. If in regard to some given event or phenomenon the women of a community say one thing and the men another, is it not because men and women have different positions in, and experience of, the same social universe of discourse?"

There would seem to be no point in asserting a collective identity unless there is an identifiable "other" that stands in opposition to it--"us" versus "them." I am interested in what the "us" consists of, and how it may influence the way individuals within that self-identified group operate. Maps--"world maps" on particular--seem to be a revealing yet rather neglected way to get at conceptions of the other, thus revealing something important about the collective self. As Sahlins remarks, "Divinities or enemies, ancestors or affines, the Others are in various ways the necessary conditions of a society's existence." A more or less "self-conscious fabrication of culture," constructed in response to "imperious outside 'pressures'" is thus a "normal" historical process.

Robert Rundstrom has observed that mapping "is fundamental to the process of lending order to the world." Yet quite clearly there are many ways of worldmaking. In Denis Wood's vivid formulation: "Every map shows this . . . but not that, and every map shows what it shows this way . . . but not the other." In other words, cartographers construct the world, they do not reproduce it. Places are where they are, but maps represent them where the mapmakers want them (or need them, or think them) to be. Every map, then, has an author, a subject and a theme (or themes). No map is a neutral document. All reflect efforts of one kind or another to impose oneself (or one's culture) on physical space. A map is an interpretation that needs, in turn, to be interpreted.

How should one go about doing this? William Boelhower emphasizes the need to take special notice of the map's hybrid nature, its "dissimulating cleverness" and "the complexity of its conventions." He urges us to feel the map's "political muscle," salute its "military potential," and delight in its "aesthetic seductiveness." In Boelhower's view, to understand the peculiar "generative logic" of maps requires a perspective that gives to the viewer a sense of the map's "spatial dynamism, its temporal narrativity, and its unfailing subjectivity." Such an interpretive stance requires the "reader" to understand how images, words and lines produce symbolic information together, how the world is transformed into these three different but related ways of "encoding space."

G. N. G. Clarke alerts us to the importance of inscriptions and other forms of "decorative" art on maps. To view cartouches as merely a source of visual pleasure, or to see them as fundamentally a distraction from the main content of a map, is to "deny the complex textuality held within the look of the map." Such a perspective, he says, "not only fails to give the map its necessary cultural status; it ignores the subtle relationship between the scientific and decorative; it fails to see them . . . as a series of interrelated indexes which bind the map within a series of ideological assumptions as to the way the land is viewed." Thus, inscriptions, colors, the presence or absence of overt symbols, even the thickness of lines, may provide clues as to the cultural purposes of maps.

Many theorists emphasize the use of cartography as a means of asserting political and social control. J. B. Harley writes, for instance: "Both in the selectivity of their content and in their signs and styles of representation, maps are a way of conceiving, articulating and structuring the human world which is biased towards, promoted by, and exerts influence upon particular sets of social relations. By accepting such premises it becomes easier to see how appropriate they are to manipulation by the powerful in society." David Harvey states more succinctly, "command over space is a fundamental and all-pervasive source of social power." As products and symbols of various kinds of authority (moral, "scientific," etc.), maps make distinctions that favor certain interests, "culturalizing the natural" through the process of identifying and naming, categorizing and containing.

Although maps are usually viewed as representations of space, they can also be taken as spaces of representation--fields of opportunity, waiting to be cultivated by acts of physical or intellectual appropriation or both. Indeed, as Boelhower points out in "Inventing America: A Model of Cartographic Semiosis," the map, as a "cultural sign," provides an "ideal text for studying the way Indian land was transformed into Euro-American territory and settlers from various nations into a homogeneous ethnos, as the ideological boast goes." . According to Boelhower, it was not so much the "discovery" of the New World that mattered as the particular way that it was seen--the sense of possibility that maps opened up. How, we might now ask, did Chinese cartographers view their craft and their world? What sort of possibilities did Chinese maps present?

From ancient times maps have served a variety of purposes in China. Many were designed as practical educational tools for scholar-officials, to guide, instruct and edify in times of both peace and war. They were also employed as a concrete means of asserting the emperor's territorial claims, whether local, empire-wide, or world-wide. Maps became symbolic tokens of exchange in China's domestic and foreign relations, and were even used to depict a perceived link between the realms of Heaven and Earth. Significantly, they also provided a means by which viewers could take "spiritual" journeys to distant lands--the cartographic equivalent of "travelling [through a landscape painting] while remaining at rest [woyou]".

Chinese mapmakers tended to be broadly gauged scholars and artists rather than narrow technicians. Until the late nineteenth century there were no professional or specialist cartographers as such in China. The scholars who created maps saw their productions as part of a larger intellectual and cultural enterprise--one that embraced not only science (especially astronomy and geography) but also history, philosophy, religion, art, literature, and religion (including divination). "History" was an especially prominent value in Chinese maps. Many cartographic collections, and even individual maps, bear titles indicating that they are concerned with the relationship between the "past and present" (gujin), or between successive dynastic periods (lidai). In other words, time and space remained closely connected in imperial China.

On the whole, explicitly religious maps seem to have been less popular in the Central Kingdom than in other parts of Asia, such as Burma, Korea, Japan and Tibet. We do, however, find Chinese cartographic works in both the Religious Daoist and Buddhist traditions. One such work, known as the Sihai Huayi zongtu (General Map of Chinese and Barbarian [Lands] within the Four Seas), purports to show the Buddhist continent of Jambudvipa (Chinese: Nanshanbuzhou), but replaces India as the principal geographical focus with China. The Chinese landscape, with its provinces, major rivers, mountains, and the Great Wall, is depicted in considerable detail, while India recedes to comparative insignificance in the southwest . A distinctive feature of this sort of map is its strong affinity with the Korean "wheel maps."

A number of Chinese maps indicate, sometimes explicitly, a concern with the principles of "siting" or "geomancy" (kanyu, fengshui, etc.). A central feature of this cosmological system is the belief that certain geographical forms and/or spaces will bring good fortune. Softly undulating rectangular shapes, for instance, are generally considered auspicious, as are lines of protective hills and mountains. Land configurations that envelope important spaces (in the fashion of the flanks of well-positioned gravesites), and waterways that nourish these areas, are also esteemed. As Philippe Foret and others have pointed out, Chinese mapmakers were not above adding such topographical features to their cartographic productions in order to depict (create) a more favorable geomantic environment. And where hills and mountains already existed but were separated by flat expanses of land that seemed to diminish their collective power, mapmakers might edit their rendering of the scene to give it greater geomantic strength. Sometimes places would simply be relocated in maps to give them a more favorable geomantic position, or altered in appearance for similar reasons.

Overall, Chinese cartographers treated large-scale space, including the world itself, as essentially flat. Although mathematical astronomers used ecliptic as well as equatorial coordinates in their celestial mapping, cartographers saw no need to project them on the earth. As a result, they "simply acted as of they were transferring points from a very large flat surface to a smaller one." At the same time, however, Chinese mapmakers often employed variable perspective and variable scale. Thus, for example, mountains might be drawn in elevation while rivers would appear in plane. Moreover, the size of objects relative to one another, as well as their distance from one another, were usually dictated not by their actual dimensions or by geometrical perspective but rather by the specific purposes for which the map was produced. Heavy annotation provided valuable information that might otherwise have been expressed by graphic images of scale.

Chinese maps often devote more space to the written text than to the actual image. Although the tendency for historians of cartography has been to denigrate heavily annotated maps in favor of more "representational" ones, there is no intrinsic reason for doing so. It was not, after all, lack of skill or "backwardness" that determined the nature of traditional Chinese cartography. In China, for cultural reasons, the written word, rather than visual images, remained the primary source of representational authority. In the pithy formulation of the well-known third century philosopher, Wang Bi, "Image is what brings out meaning; word is what clarifies image."

Cartographic texts in China commonly provided technical data concerning roads, waterways, landmarks, distances, and so forth. But they also supplied important cultural information. An excellent illustration can be found in a "geographic map [of China]" (Zhuili [Dili] tu), created by a scholar named Huang Chang in the 1190s--several decades after the fall of the Northern Song capital of Kaifeng to the invading Ruzhen people. The map was intended as an illustration for the future Song emperor (Ningzong, r. 1194-1224) of how much land had been lost to the northern barbarians, and as a reminder of the sovereign's responsibility to reunite the empire. The commentary to the map addresses the perennial problem of keeping China together, observing that "only one out of every ten [rulers] has been able to bring unity to all under Heaven." This discussion--which constantly emphasizes morality as the key to administrative success--is full of historical allusions to events such as the invasion of China by the Qidan people in the early tenth century and the rebellion of An Lushan in the mid-eighth century, as well as to the noteworthy unifying accomplishments of the sage-rulers Tang and Wen, who, despite having only modest territories to begin with (like the Southern Song), founded the great Shang and Zhou dynasties, respectively. The commentary naturally includes references to northern landmarks that had recently fallen under "barbarian" control--including the Yellow River, the Great Wall, and "a vast forest stretching several thousands of li."

The textual emphasis of traditional Chinese cartography did not in any way undermine the aesthetic appeal of maps. On the contrary, inscriptions often enhanced it. In contrast to the development of cartography in Europe, where manuscript maps became rather rare following the spread of copper engraving in the late fifteenth century, manuscript maps continued to be produced in great numbers in China. These documents, like landscape paintings, were tastefully shaded and often complemented by substantial amounts of calligraphy--sometimes even poetry. Printed maps could also be extraordinarily beautiful, with handsome, well-cut cartouches, and carefully colored natural features. Neither type of map could be considered true art, however, for both lacked the qualities of "life force" (qi) and "kinesthetic power" (shi) that indistinguished artistic creativity from mere craftsmanship.

Yet another distinctive feature of Chinese cartography is what Cordell Yee describes as its tendency toward introspection--a self-conscious preoccupation with concrete administrative concerns. Buildings and walls, for example, tend to loom large, quite literally, in many kinds of maps. Paradoxically, Chinese "introspection" included looking outward. That is, one of the emperor's traditional "domestic" concerns as the ruler of "all under Heaven" was the management of foreign peoples--whether on the periphery of his realm or beyond. These "barbarians" (yi, fan, etc.), although by definition not fully Chinese, were all at least theoretically the emperor's "subjects." Many of them periodically sent him local products, designated "tribute" (gong), and, in return, expected the Son of Heaven to protect and nurture them. From a Chinese standpoint, this highly refined system of "guest ritual" (binli), which allowed foreigners the opportunity to demonstrate their loyalty to the Chinese emperor, was the logical extension of an ancient "feudal" structure of lord-vassal relationships. Although the tributary system underwent many permutations over time, what remained constant was a highly refined vocabulary of imperial condescension that at once emphasized the inferiority and encouraged the loyalty of all China's tributaries, far and near. It was this Sinocentric assumption of universalistic overlordship--the idea of a Chinese "empire without neighbors"--that blurred the distinction between maps of "China" and Chinese maps of "the world."

The earliest extant "world-maps" in China date from the Song dynasty (960-1279). One example is the Gujin Huayi quyu zongyao tu (General Map of Chinese and Barbarian Territories, Past and Present), which dates from about 1100. Another closely related and far more famous example is the Huayi tu (Map of China and the Barbarians; 1136). This latter work, about three-foot square and carved in stone, supplies approximately 500 place names and identifies a dozen or so rivers and tributaries in China. A few foreign lands are represented visually in the map--notably, Korea and India--but more than a hundred different groups of "barbarian" peoples are indicated only by written notes on the margins of the maNear the top, on the northwestern side we learn, for instance, that the area of the formerly enfeoffed Qidan people "is now called 'the Great Liao Country." Several such notes refer specifically to tributary relationships, past and present.

Not all Song dynasty renderings of space arose from the same source, however. Indeed, inscribed on the reverse side of the Huayi tu is an astonishingly "modern" looking version of an ancient work called the Yuji tu (Map of the Tracks of Yu, 1136), probably created about 1080. It marks the earliest extant example of the so-called "latticework" cartographic grid in China. Each side of each square represents 100 li (c. 33 miles), yielding a scale of about 1:1,500,000. The outstanding feature of this map, in addition to the near total absence of written commentary, is its extremely "accurate" depiction of major landforms. The representation of China's coastline, for instance, looks remarkably like modern twentieth century renderings.

The grid system of the Yuji tu apparently provided the model for Zhu Siben's influential Yutu (Terrestrial Map; c. 1320). The only extant version of this work is Luo Hongxian's Guang Yutu (Enlargement of the Terrestrial Map), first published in 1579. Luo's production takes the form of an atlas, with more than forty separate maps--including a "General Map [of China]" (Yudi zongtu) and a "General Map of China and the Barbarians" (Huayi zongtu). Like the Yutu, the Guang Yutu employs a grid system, but unlike Zhu's map, Luo includes a number of cartographic legends--twenty-four in all--for mountains, rivers, boundaries, roads, and other landmarks.

Luo's atlas obviously reflects, not least in its abundant written texts, the expansion of Chinese knowledge about the rest of the world gained in the course of the eunuch-admiral Zheng He's extensive naval expeditions during the early fifteenth century--voyages which took him as far west as the shores of east Africa. As one measure of its comprehensive scope, the Guang Yutu includes an elaborate chart that distinguishes the residents of over 120 foreign countries by area: Eastern Barbarians (Koreans and Japanese), Southeastern Barbarians (Liuqiu Islanders), Southern Barbarians (Southeast Asians), Southwestern Barbarians (Filipinos, Indians, Westerners, etc.), Barbarians of the "Western Regions" (including various Turkic peoples) and Northwestern Barbarians (Mongols and other such tribes). Many of these peoples are designated "tributaries," not only in the chart but also on some of the maps themselves.

Luo's work spawned a number of imitations, including the Da Ming guangyu kao (An Examination of the Enlarged Terrestrial [Map] of the Great Ming Dynasty; 1610) and Chen Zushou's Huang Ming zhifang ditu (An Administrative Map of the Ming Dynasty; 1636), banned during the Qing period. It is important to remember, however, that most large-scale Chinese maps of the late imperial era continued to conform to the gridless Huayi tu cartographic model. (478) The most striking and expansive example from the early Ming period is the magnificent, multicolored Da Ming hunyi tu (Amalgamated map of the great Ming empire; c. 1390). Drawn on a horizontal scale of 1:820,000 and a vertical scale of 1:1,060,000, it covers an area extending all the way from Japan to the Atlantic Ocean (including both Europe and Africa), and from Mongolia to Java. Although the section on China seems to be derived primarily from Zhu Siben's Yutu, the renderings of Africa, Europe and Southeast Asia appear to have been based at least in part on Li Zemin's Shengjiao guangbei tu (Map of the Vast Reach of [China's Moral] Teaching; c. 1330), no longer extant."

Subsequent maps based on the Da Ming hunyi tu model tended to be somewhat more restricted in geographical scope, but still impressive in their coverage. The best sixteenth century example is Yu Shi's Gujin xingsheng zhi tu (Map of Advantageous Terrain, Past and Present; 1555). This beautifully colored and heavily annotated work--representing an expanse of territory stretching from Samarkand, India and Arabia in the west to Japan in the east, and from present-day Mongolia in the north to Java and Sumatra in the south--lacks any sort of grid. It elongates Korea, treats the Shandong penninsula as if it were an island, and, like the Huang Ming yitong dili zhi tu, refers to the existence of several mythical places derived from the ancient Shanhai jing (Classic of Mountains and Seas). This work--the earliest Chinese illustrated account of "barbarians"--describes a great number of foreign lands with all sorts of exotic inhabitants: societies consisting of women, or giants, or dwarfs; people with multiple heads or bodies; creatures with the heads of humans and the bodies of snakes; and so forth.

A horizontally oriented version of Yu's map, titled Gujin tianxia xingsheng zhi tu (Map of the Advantageous Terrain under Heaven, Past and Present; n.d.). appears in Zhang Huang's Tushu bian (Compilation of Illustrations and Writings; 1613). Like Yu's production, it is full of historical references, including information on the activities of China's "barbarian" neighbors, the development of the Chinese tributary system, and various administrative changes within the Chinese empire. To an even greater extent than Yu's map, the Gujin tianxia xingsheng zhi tu identifies the homes and/or exploits of China's great culture heros, ranging from Confucius and his followers, to the Tang poet, Li Bai, to the founding emperors of the Tang, Song and Ming dynasties. It also refers to the activities of several prominent Chinese loyalists, including Zhuge Liang of the Three Kingdoms period and both Yue Fei and Wen Tianxiang of the Song.

Joseph Needham argues that there was a general "advance" in Chinese mathematical cartography from the Song period into the seventeenth century. In fact, however, the evolution of map-making in China cannot be characterized as simply a linear process of "progressive" improvement. Rather, Chinese cartographers continued to produce two distinctly different types of maps--one based on relatively precise mathematical measurements, and one based primarily on cultural data--without explicitly recognizing the existence of two competing traditions. If a characterization is required, it would have to be that maps of the latter sort greatly outnumbered those based on more mathematical models--not only up to the seventeenth century but well beyond. On the other hand, as we shall see, a number of cartographic documents of the late Ming and Qing periods placed the two types of maps together, in the spirit, one suspects, of the complementary maps engraved on the two sides of the Song stele of 1136.

Many modern scholars, both Western and Chinese, have seen the arrival of the Jesuits in China during the late sixteenth century as a landmark in the history of Chinese map-making. In fact, however, their influence was rather limited. To be sure, Jesuit scientific methods, including sophisticated surveying techniques, enabled the Qing dynasty (1644-1912) to create a far more mathematically "accurate" map of the Chinese empire than had ever been produced before--the Huangyu quanlan tu (Map of a Comprehensive View of Imperial Territory; 1718). This massive work, the product of many years of dedication by both the Jesuits and Qing scholars, provided China's Manchu rulers with an important instrument of political and military control, and it remained the most authoritative atlas of the realm for nearly two centuries.

But from the standpoint of world maps, Jesuit mappaemundi--including Matteo Ricci's Yudi shanhai quantu (Complete Map of the Earth's Mountains and Seas; 1584), his Kunyu wanguo quantu (A Complete Map of the Myriad Countries of the World; 1602), Giulio Aleni's Zhifang waiji (Notes on [World] Geography, 1623), and Ferdinand Verbiest's Kunyu tushuo (Illustrated Discussion of the Geography of the Earth; 1674)--had little long-term influence. Whereas precise maps of the empire had obvious strategic value, especially for the expansive but alien and somewhat insecure Manchus, world maps had a different function altogether. They were designed primarily as visual statements about a great and glorious culture, a universal order focused squarely on the Chinese tributary system. Indeed, one gains the impression that most Chinese world maps were constructed as if they were to be seen by the emperor himself.

It is not surprising, then, to find that a number of Chinese scholars bitterly attacked the Jesuits for misrepresenting the world and China's place in it. According to one Ming scholar, Wei Jun, Ricci's map not only contained "fabulous and mysterious" information that could not be verified, but in locating China to the west of center and inclined to the north, it dislodged the "Central Kingdom" from its rightful position at "the center of the world." How, Wei asked, "can China be treated like a small unimportant country?" Similarly, the Huangchao wenxian tongkao (The Imperial Dynasty's Comprehensive Examination of Source Materials; 1787) denounced Ricci's account of the world as full of contradictions, misguided statements and "boastful lies" (dankuang). It accused him of belittling China, aggrandizing his own culture, and spreading misinformation in the course of his cartographic work.

Even individuals who claimed to have been directly inspired by the Jesuits often borrowed little of cartographic substance from them. One noteworthy example is a map by the scholar-official Liang Zhou, titled Qiankun wanguo quantu gujin renwu shiji (Universal Map of the Myriad Countries of the World, with Traces of Human Events, Past and Present; c. 1600). This work--which appears to have been created more out of defiance than admiration--bears no trace of meridians and arranges foreign locations topologically rather than topographically. About eighty transoceanic lands outside of China appear in this form, in addition to a hundred or so additional foreign places to the north and west. Locations such as North America (on the upper right-hand side of the map) and South America (on the lower right-hand side)--like the Land of Tall People, the Land of Small People, the Land of Women, and many other places drawn directly from the pages of the Shanhai jing--are shown as inconsequential islands surrounding the large nucleus of the Chinese empire.

One might well think that scholars of "empirical research" (kaozheng xue) in the seventeenth and eighteenth centuries would appreciate Jesuit learning, and indeed many did; but most Chinese intellectuals drew quite selectively from the available pool of Western scientific knowledge. Ironically, a deep distrust of symmetry and regularity on the part of kaozheng scholars hostile to traditional cosmography led them to reject the notion of a lawful, uniform, and mathematically predictable universe. Thus, for instance, the great Qing intellectual, Wang Fuzhi, dismissed the round-earth concept of the Jesuits out of hand.

Meanwhile, a turn inward in Chinese thought after 1644 diverted attention away from Jesuit-style conceptions of the external world. Gu Yanwu, a towering figure in early Qing scholarship, makes no mention of Jesuit world maps in his otherwise comprehensive Tianxia junguo libing shu (Treatise on the Advantages and Disadvantages of the Commandaries and States of the Empire; 1662). This lack of a serious interest in the Western world encouraged Gu to describe Portugal (Fulangji) as simply a one-time tributary state, located "south of Java," whose early contact with China was for the purpose of studying trade routes and "buying small children to cook and eat."

Of course, cartographic decisions do not necessarily involve either/or choices. In fact, a spirit of compromise animated a number of Chinese map-makers in late imperial times. Beginning in the waning years of the Ming dynasty, Chinese scholars tried various techniques designed to unite Jesuit-inspired knowledge and more traditional Chinese cartographic renderings of space. An excellent example can be found in Cao Junyi's ambitiously titled Tianxia jiubian fenye renji lucheng quantu (A Complete Map of Allotted Fields, Human Events and Travel Routes [Within and Without] the Nine Borders Under Heaven; 1644). This handsome cartographic document, which continued to serve as a model for cartographers during the Qing period, acknowledges the existence of Europe, Africa, the Middle East and India, but the two latter areas are represented primarily by cartouches, and Africa--which appears only about one-tenth the size of China--hangs down on the west side of Cao's map as if it were little more than a protective flank. Europe, tiny and even more marginal, is barely recognizable in the upper northwest portion of the maMost of the place names in these distant areas have been derived from Jesuit sources; but in the southeastern seas there are a number of mythical countries taken directly from the Shanhai jing.

The map gestures toward mathematical accuracy by providing longitudinal lines and degrees, and by supplying the estimated distances of various "barbarian" countries from the southern Ming capital (modern-day Nanjing). Moreover, in his extensive written commentaries, Cao provides a great deal of solid administrative data and historically grounded information on China's strategic rivers, lakes, mountains and seas. At the same time, however, he is pains to locate his discussions of world geography within the traditional confines of both the Chinese tributary system and Chinese cosmology. Furthermore, in his discussions of "barbarians" he does not differentiate clearly between actual foreign countries and the lands and peoples described in the Shanhai jing. The general Sinocentric spirit of Cao's map is captured in the remarks of his contemporary, the cartographer Chen Zushou: "All the barbarian people within the Four Seas should come to pay tribute to the Chinese Emperor. Although they [the Jesuits] might describe the world as comprising Five Continents, yet four of them should surround the nucleus of China."

Another kind of cartographic compromise appears in the form of a large anonymous scroll known as the Sancai yiguan tu (Illustrations of the Unity of the Three Powers [Heaven, Earth, and Man]; 1722), archived in the British Library. Although this document consists primarily of a written text dealing with history, morality, cosmology and military affairs, it includes two red planispheres, a "Comprehensive Map of Heaven and Earth," and a "Perpetual Map of the Unified Qing empire." The former map is quite clearly based on Jesuit cartography, while the latter seems to represent an unusual amalgamation of the Song Huayi tu and Yuji tu traditions.

A similar approach appears in a map produced by Ma Junliang, a 1761 jinshi degree-holder who was well-known for his skill as a mapmaker. In the 1780s or early 90s, Ma produced a large and widely distributed woodblock print titled Jingban tianwen quantu (Capital Edition of a Complete Map [of the World Based on] Astronomy), which featured a traditional-style rendering of "the world" based more or less on the time-honored model of Liang Zhou. But Ma also offered on the same sheet of paper a pair of global maps--one derived from a loose rendering of Matteo Ricci's mappamundi that appears in the Ming encyclopedia Sancai tuhui (Illustrated Compilation of the Three Powers; c. 1607) and one borrowed from a similarly structured Chinese map of the eastern hemisphere, first published by Chen Lunjiong in his Haiguo wenjian lu (Record of Things Heard and Seen in the Maritime Countries; 1730). [For an on-line image of this map, the original of which is in the Woodson Research Center of Fondren Library, see http://www.ruf.rice.edu/~asia/jingban.html]

Ma's production may be considered an innovative offshoot of a genre of "complete maps of all under Heaven" (Tianxia quantu) that arose in the late seventeenth century and seems to have dominated Chinese visual representations of "the world" until at least the mid-nineteenth century. Although these maps all conformed to the basic Liang Zhou model, they also included cartographic symbols of the sort found in the Guang Yutu, and easily accomodated variations such as those in Ma's Jingban tianwen quantu. Different editions of these attractive, often delicately tinted maps were produced by a series of Qing scholars, including Ma, at least two predecessors--Yan Yong (fl. c. 1710) and Huang Qianren (fl. c. 1770)--and at least one successor, Zhu Xiling (fl. c. 1820). Of these individuals, only Yan is known to have employed a grid system, following the lead of the great Qing scholar, Huang Zongxi, who produced a rather stark "map of China" (Zhongguo ditu) in 1673. Later versions of the Tianxia quantu genre--at least the ones that I have seen--lack any traces of a grid, although they generally acknowledge a direct cartographic debt to Huang Zongxi (Qianren's grandfather).

Most maps of this variety are known by the title Da Qing yitong tianxia quantu (Complete Map of the Comprehensive Great Qing Empire) or a close equivalent. Some, however, bear significantly different names as a way of highlighting certain additions to the basic cartographic format. For example, in addition to Ma Junliang's Jingban tianwen quantu we find an anonymous work titled Jingban tiandi quantu (Capital Edition of a Complete Map of Heaven and Earth), which includes a round star chart above the standard terrestrial image. There are also certain minor differences in quality, color, commentaries and a few place names in Tianxia maps.

Some discrepancies appear to be simple scribal mistakes, such as writing "ten thousand li" instead of "twenty thousand li" (the usual figure) for the extent of the Russian empire; or confusing Zhu Siben's family name with another, similar-looking character. Others involve the expansion or contraction of information--the inclusion or omission of a certain source of authority, or varying degrees of specificity regarding time periods and other minor details. Once in a while there is a major discrepancy. For instance, on some maps the characters "Small Western Ocean" appear where the characters "Great Western Ocean" would be expected. Occasionally, delicate coloring gives way to much darker and less attractive tones. On the whole, however, the similarities are far more striking than the differences.

Like virtually all large-scale Chinese maps, works of the Tianxia quantu genre convey a vivid sense of China's vast and varied landscape: its mountain ranges, overland travel routes, river systems, lakes, coastal communications, and deserts (particularly the Gobi), as well as other prominent landmarks--notably the Great Wall and the "Sea of Constellations" (Xingxu hai)--the legendary source of the Yellow River. An intriguing feature of every Tianxia map I have seen is a prominent stone tablet (bei) erected to Zhuge Liang in the far southwest, presumably out of appreciation for his role in pacifying China's borders during the Three Kingdoms period (222-265). This is the only example of an individual so honored in these maps; even the birthplace of Confucius at Qufu has only a general reference to the Kong family graveyard (Konglin), not a specific tablet.

As with works of the Yu Shi tradition and those undertaken with Jesuit assistance in the early eighteenth century, maps of the Tianxia quantu variety pay close and careful attention to administrative changes within China's provincial boundaries--changes resulting from divisions, consolidations and other adjustments. All maps of this sort employ sets of eight or so cartographic symbols to indicate important administrative centers such as provincial capitals (sheng), prefectures (fu), departments (zhou), districts (xian), passes (guan), garrison towns (yingzhen) and so forth. They also mark the presence of local headmen (tusi)--members of ethnic minorities responsible for supervising their own people and for reporting periodically on them to regular Qing officials.

Of particular importance to the authors of Tianxia maps are changes in the size and shape of the Chinese empire occasioned by Qing military conquests during the seventeenth and eighteenth centuries. The introduction to every new edition of these maps begins with exactly the same proud phrase: "The land ruled by the present dynasty is unprecedented in its extent." Each cartographer then proceeds to describe the specific political and military developments that made a new map necessary. For the most part, these developments had to do with campaigns by either the Kangxi or the Qianlong emperor which brought large areas of the Inner Asian frontier under direct Qing control. But they also involved negotiations between the Chinese and representatives of various "outer" peoples, from Tibetans to Europeans.

Significantly, "outer" areas are not at all well defined in maps of the Tianxia quantu genre. Representationally speaking, there are no ovious borders separating China from Russia or clearly delineating the individual kingdoms and territories of, say, Central Asia, India, or mainland Southeast Asia. In a few cases boundaries are suggested by written inscriptions; but only the oceans and seas allow certain countries to appear fully detached from the Central Kingdom. On three sides of the mainland such places are represented as islands, with written descriptions of varying lengths. There is no effort to show the relative size of foreign realms pictorially, however. Rather, following a pattern established by the Huayi tu and continued by such cartographers as Yu Shi and Liang Zhou, the size of foreign territories often appears to be a function of the amount of text deemed necessary to tell the viewer what needs to be known about them. (Unlike these works, however, Tianxi maps do not refer to mythical lands from the Shanhai jing). Most inscriptions provide useful historical background on the "barbarians" in question, including place-name changes and changes in their relationship to China over time. Sometimes they also supply data on distances, including travel routes and the length of an occasional border.

We should not think, however, that cartographers working in the Tianxia quantu tradition were unconcerned with representing foreign lands and peoples as accurately as possible. In a lengthy introduction to his map of 1714, Yan Yong candidly admits that limitations of both information and cartographic space prevent him from showing the actual locations of far away places. Nonetheless, he has tried to indicate their relative positions and to include textual information on their approximate distance from China. Although most later maps of the Tianxia quantu genre do not bother to make this point or to offer systematic data on distances, Ma Junliang's Jingban tianwen quantu offers an interesting solution to the problem, as we have seen. By combining two radically different types of maps in one document, he gives his viewers an epistemological choice. Rather than trying to reconcil the two versions, Ma leaves the issue open--a cartographic compromise reminiscent of the Song dynasty stele bearing the two radically different maps of 1136.

From the Song dynasty onward, virtually all large-scale maps that deal with "barbarians" of any sort refer to the Chinese tributary system. Works of the Tianxia quantu genre are certainly no exception. Invariably they include textual information on the background and evolution of important Sino-foreign tributary relationships, the frequency of certain missions, and major tributary routes. In so doing they reveal a rich lexicon of tributary terminology. The preface to each map refers explicitly to the process by which barbarian envoys come to China and offer themselves as vassals of the Qing dynasty. This process of symbolic submission is always described as an arduous one, involving "the scaling [of mountains], the sailing [of seas], and several stages of translation [ti hang chongyi]."

Such cartographic clichés reflect deep-seated attitudes expressed in a number of official documents, including the Huang Qing zhigong tu (Illustrations of the Tribute-Bearing People of the Imperial Qing; 1761), the Huang Qing fanbu yaolue (Essentials of the Vassal [Tribes] of the Imperial Qing; 1845), and the Chouban yiwu shimo (Management of Barbarian Affairs from Beginning to End; 1880). The prefaces to each of these compilations display the same condescending tone. The first emphasizes how "within and without the empire united under our dynasty, the barbarian tribes have submitted their allegiance and turned toward [Chinese] civilization [xianghua]." The second, by the great Qing geographer, Li Zhaoluo, refers to the way the emperor "nourishes [his dependencies] like their father and their mother," and "illuminates them like the sun and the moon." And the third, using much the same language as the first, describes the historic process by which foreigners gravitate to China, become "cultivated" and learn "elegance and etiquette."

The ten volumes of the Huang Qing zhigong tu provide a detailed picture of the Qing tributary system in its heyday. Most of these volumes deal with the peoples of Inner Asia and the ethnic minorities of Southwest China. Yhe first, however, focuses on China's overseas tributaries, listed in the standard order: Korea, the Liuqiu Islands, Annam, Siam, Sulu, Laos, Burma, and the Great Western Ocean (Da Xiyang). These discussions are followed by sections on the Small Western Ocean (Xiao Xiyang), England, France, Sweden, Holland, Russia, and the Phillipines.

Here, without benefit of cartographic representation, the so-called Great Western Ocean Country is located vaguely in the Atlantic region and identified both with Italy and Portugal. Other Western nations, including England, France, Sweden, Holland, and Russia are lumped together indiscriminately with Asian countries such as Japan, Borneo, Cambodia, Java, and Sumatra. Modern France is confused with Ming dynasty Portugal; and England and Sweden are recorded as countries dependent on Holland. According to the Huang Qing zhigong tu, Italy presented tribute in 1667 (it was actually Holland that did so) and the Pope himself is reported to have once brought tribute to China. In religious matters the Huang Qing zhigong tu informs us that the Portuguese/French were Buddhist countries before they accepted Catholicism.

The same kind of misinformation can be found in the section on "tributary states" in various editions of the Da Qing huidian (Collected Statutes of the Great Qing Dynasty). Thus we read in the Collected Statutes of the Jiaqing reign (1796-1820) that "Portugal [Gansila] is in the northwestern sea near England," and that "France [Falanxi], also called Fulangxi, is the same as Portugal [here, Folangji]." After absorbing the Phillipines (Lusong), this account goes on to say, "they [the French/Portuguese] divided their people and lived there, still governing it at a distance. . . . The sea route from this country to China is more than 50,000 li [c. 17,000 miles]." Sweden (Ruikuo), we are told, is in the northwestern sea; the distance by sea is calculated to be over 60,000 li [c. 20,000 miles]. [ . . . ] Denmark [Lianguo] is [also] in the northwestern sea, and its route to Guangdong province is the same as that for Sweden." Small wonder, then, that Chinese mapmakers found it difficult to acquire accurate data on foreigners. In fact, some of the same misinformation cited above is repeated in maps of the Tianxia quantu variety--for example, that the Portuguese (Gansila) "absorbed" the Phillipines.

More reliable information was, however, available, as we can see from a large, hand-colored cartographic scroll produced by a scholar named Zhuang Tingfu in 1794. The title of Zhuang's production is: Da Qing tongshu zhigong wanguo jingwei diqiu shi (Model of the myriad tributary states of the great Qing dynasty from around the globe). Although this work borrows certain cartographic elements from Ma Junliang's Jingban tianwen quantu--specifically, the Sancai tuhui version of Ricci's map and Chen Lunjiong's depiction of the eastern hemisphere--it pointedly ignores the lower part of Ma's map. Instead, it provides two extremely "modern"-looking renderings of the Eastern and Western hemispheres, both produced by Zhuang himself. These latter two maps were reprinted by Western-oriented Korean exponents of "practical learning" during the 1830s.

A pair of long written inscriptions, totalling about five thousand characters, illustrate Zhuang's two major themes: one, the transmission to China of new Western scientific knowledge by the Jesuits; the other, the historic process by which foreigners came to be ruled (laiwang) as vassals; that is, they "knocked on [China's] gates," "sincerely offered tribute," and asked to become "attached" (shu) to the Central Kingdom. The interesting feature of Zhuang's document is the way it accomodates simultaneously the idea of embracing new knowledge from the West and the notion of enrolling Westerners as traditional-style tributaries.

From a scientific standpoint, Zhuang seeks to show that he has learned a great deal from the Jesuits about geography, cartography and astronomy, which, indeed, he did. He waxes at length about latitude and longitude, time and seasonal change, the circumference of the earth (90,000 li), the north and south poles, and so forth. He also writes knowledgeably about how different cartographic projections yield different pictures of the world. According to Zhuang, previous maps, including those offered by the Jesuit fathers, Matteo Ricci and Ferdinand Verbiest, distorted China's size by placing it too far north, thus compressing it (making China appear too small and the foreign countries, too big). His map, drawing upon the work of the famous Qing scientist Mei Wending, provides, he says, a more accurate picture. Significantly, Zhuang cannot resist remarking on how, cosmologically speaking, the Chinese are fortunate to have been born in the Central Land (Zhongtu), where the radiance of the sun nourishes them like a sovereign or a father--unlike those people whose misfortune it is to be in far northern or southern regions, where beneficial qi is less direct and therefore not very helpful.

Although Zhuang devotes a great deal of attention to science, his primary concern is a cultural one: the Chinese tributary system. The Da Qing tongshu zhigong wanguo jingwei diqiu shi commemorates the well-known Macartney embassy of 1793, which, in turn, marked what Zhuang considers to be the highwater mark in the development of China's age-old system of hierarchical foreign relations. This system, he notes, expanded significantly during the Kangxi and Qianlong reigns to include many new parts of the "Western Regions." The peoples of these areas, Zhuang goes on to say, have been registered as part of the Chinese empire [ru banji], and have offered tribute to the Qing dynasty along with the British, who had engaged in no official communication with China prior to 1793.

Earlier maps, Zhuang tells us, did not include all of China's tributaries; but the Macartney mission, together with the "coming to court" of other tributaries, and the "return" (laigui) of various tributary peoples from the "Western regions" during the eighteenth century, offers a fitting moment to celebrate the transformative effect of the throne's glory (shenghua) with a set of maps. His renderings, then, are respectfully offered on this magnificent occasion. Significantly, but not at all surprisingly, Zhuang's remarks about the civilizing role of the Chinese emperor (shengjiao) correspond closely to those provided in the major cartouche of Ma Junliang's Jingban tianwen quantu.

The late eighteenth century maps by Ma, Zhuang, and others bring into sharp focus the issue of how best to characterize the Qing tributary system. James Hevia's stimulating book, Cherishing Men from Afar (1995), which deals with the Macartney embassy, emphasizes the flexibility of the Chinese system, and argues that Qing guest ritual "does not appear to deal in crude distinctions between civilization and barbarism." Although Qing officials and the throne did indeed evince a good deal of flexibility in "managing" Macartney (part of a long-standing tradition), and although the word "crude" probably does not apply to their approach, I believe that Hevia underestimates the problem of cultural difference. What distinguished the "Chinese" from "barbarians" was precisely the difference in their levels of "civilization"--specifically, differences in their ritual behavior.

There is another problem with Hevia's approach. Although his stated aim is to understand events "through their multiple recountings," his analysis is marked by a curious asymmetry. In his zeal to expose the "orientalizing" tendencies of both Westerners and post-Qing Chinese scholars (who have, according to Hevia, appropriated "the intellectual framework of the colonizer"), he virtually ignores similar "occidentalizing" gestures on the part of the Qing intelligentsia--essentializing and condescending moves that are abundantly evident not only in the Chinese documents that Hevia has quite obviously studied, but also in Chinese cartographic materials, which he apparently has not. The result is an account of historiographic "distortions" that is itself "distorted" by Hevia's inclination to view precolonial China through a postcolonial lens.

My own research on China's foreign relations, including recent work on maps and other forms of visual representation, suggests that in late imperial times there was something that could indeed be called a Chinese tributary "system" in Qing times, universally recognized by all Han people, marked by the generic term zhigong (signifying the offering of "regular tribute"), and expressed in a wide range of elite as well as popular writings and illustrations. This system, although by no means the only mechanism for the conduct of Sino-foreign relations, was highly sophisticated, remarkably flexible, and perfectly "rational"--particularly in the light of Chinese cultural assumptions about imperial overlordship, the transformative effect and power of ritual, and the "nature" of both foreigners and Chinese. These assumptions, I might add, seem just as "totalizing" and demeaning with respect to foreigners ("barbarians") as those of the British, which Hevia catalogues at far greater length. In fact, references to the animal-like qualities of Westerners abound in Chinese writings, official and unofficial as well.

Qing documents from the last Dutch embassy to China in 1794-1795, a year or so after the Macartney episode, help to explain why the Chinese tributary viewpoint proved to be so tenacious, and why the Qing authorities were so confused and dismayed by Great Britain's failure to meet Chinese expectations in 1793. Although the Dutch mission, designed to celebrate the Qianlong emperor's sixtieth birthday (and, of course, to promote trade), was considered an irregular event, it conformed in every respect to the basic requirements of Chinese tributary ritual. Holland's preliminary letter, sent from its fictitious "king" to the Qianlong emperor in the summer of 1794, captures the distinctive flavor of a vassal's petition to his feudal superior. The Chinese version reads in part: "From the time of the Kangxi emperor's reign [1662-1722] onward . . . [we foreigners] have all been transformed by China's civilizing influence [xianghua]. Throughout history there has never been a monarch with such a peerless reputation as you possess, my exalted emperor."

Small wonder China had such a well-developed sense of its exalted status. In response to the Dutch mission, the Qianlong emperor replied:

I have now reigned for sixty years, so that the four seas are forever pure and all the regions of the world have all been transformed by Chinese culture. My virtuous reputation has spread far and wide . . . and I have [always] treated Chinese and foreigners as one family. . . . Now [representatives of] the myriad countries, scaling mountains and sailing seas, have come, one after another, to offer birthday congratulations. . . . [This] heavenly dynasty views all [people of the world] with equal benevolence, and although some may come to China with only meager [tributary presents], all will leave amply rewarded . . . . [Since you admire Chinese culture (muhua) and will be receiving valuable tokens of imperial favor with this edict,] may these gifts strengthen your bonds of loyalty and integrity, preserving good government in your kingdom and making you forever worthy of my esteem."

As I have argued elsewhere, well into the nineteenth century Qing documents on foreign relations, including maps, continued to reflect this tone, using the same basic vocabulary.

It simply will not do to dismiss this pervasive and tenacious world view as empty rhetoric. One can acknowledge the flexibility and sophistication of traditional China's approach to foreign relations--and even accept James Polachek's argument that the "inertia of the [Chinese] central-government political system . . . [was] the chief obstacle to foreign policy change"--without denying, as Polachek himself puts the matter (somewhat indelicately), "the pompous 'Celestialism' of the late Ch'ing [Qing] court posed a very real problem for those would have brought China more speedily into the modern world."

This much we know: On the eve of the first Anglo-Chinese War of 1839-42, world maps of the sort produced by individuals such as Cao Junyi and Zhuang Tingfu were at best a dim memory for most Chinese scholars. From the late seventeenth century into the nineteenth, the vast majority of Chinese mapmakers ignored Jesuit constructions of the world almost entirely. Most did not even chose to pattern their cartographic productions after Luo Guangxian's grid-oriented Guang Yutu. Far more popular were maps of the Tianxia quantu variety, or those based on the rhythmic and colorful cartography of Yu Shi. A striking example of the latter type is an untitled and anonymous "world map" of 1743 that is now housed in the Oriental Manuscripts Division of the British Library.

Yet at least a few indigenous mapmakers carried on the cartographic traditions established by the Jesuits--assisted now by the efforts of a new breed of Western missionaries, primarily Protestants. For instance, Li Mingche, a well-known Daoist priest and scientist with foreign contacts, included two relatively "modern" illustrations of the Eastern and Western hemispheres--complete with lines of latitude and longitude--in his Huantian tushuo (Illustrations of Encompassing Heaven; 1819). Moreover, every late eighteenth and early nineteenth century example I have seen of the defensively oriented scrolls known generically as Haijiang yangjie xingshi quantu (Complete Map of [China's] Coastal Configurations) begins with a colorful and quite faithful line-drawn rendering of the eastern hemisphere based on Chen Lunjiong's Haiguo wenjian lu--the same basic model that appears on Ma Junliang's maps in the Tianxia quantu tradition. Although the Chinese empire usually appears in such renderings to be as large as Africa (which embraces one-fifth of the earth's total land area), the sailient feature of the map is its unmistakably "modern" appearance.

After China's defeat at the hands of the British in 1842, editions of the Haijiang yangjie xingshi quantu begin to reflect a new awareness of the Western presence in treaty port areas--significantly, by means of textual remarks to the effect that "during the Daoguang period [1820-1850] Western countries [began to] trade at this place." Such cartographic changes were part of a growing sense on the part of at least some Chinese scholar-officials that China had to acquire more up-to-date knowledge about the outside world in order to survive. The two most important books to provide this information, both based substantially on Western maps, were Wei Yuan's Haiguo tuzhi (Illustrated Gazetteer of the Maritime Countries; 1843) and Xu Jiyu's Yinghuan zhilue (A Short Account of the Maritime Circuit; 1849). What these two pioneering works had in common was a desire to, in Wei's words, "describe the West as it appears to Westerners."

From the 1860s to the 90s, as part of China's Self-Strengthening Movement (1862-1895), study associations, books, and journals devoted to geographic and cartographic issues began to proliferate in China. The publication of Wang Xiqi's massive Xiaofanghu zhai yudi congchao (Collected Texts on Geography from the Small Square Vessel Studio; 1877-1897), which brought together several hundred individual Qing dynasty titles, marked a watershed in China's geographical awareness. Meanwhile, Chinese cartographers began to produce their own colorful, modern-looking maps. The Beijing Library has collected several of these works, with titles such as Diqiu wu tazhou quantu (Complete Map of the Five Great Continents of the Globe; 1874); Diqiu quantu (Complete Map of the Globe; 1883) and Diqiu wanguo quantu (Complete Map of the Myriad Countries of the Globe; 1895. Certain mapmakers, including Yao Wentong, Gong Zhai, Chen Zhaotong and Wen Shao, even managed to achieve a certain limited celebrity. In the 1880s and 90s, the Qing court itself attempted to update and standardize its geographic and cartographic practices.

But it was the Sino-Japanese War of 1894-95 that sounded the death knell of traditional Chinese cartography. From this time onward, in elite journals and even popular almanacs and encyclopedias, Chinese readers sought ever more accurate knowledge about other parts of the world, including once-despised Japan. The rise of Chinese nationalism--generated by China's humiliating defeat at the hands of the so-called "dwarf-bandits" (Wokou)--brought a heightened awareness of the wages of foreign imperialism. Chinese cartography, like many other areas of Chinese life after 1895, became suffused with the spirit of patriotism and political action. One revealing illustration can be found in a map contained in a 1912 almanac (Zhonghua minguo yuannian lishu), issued in the name of the newly established Republic of China. Although not particularly sophisticated in terms of mathematical cartography, the map is fascinating because its commentaries strikingly identify the various places taken from China by foreign imperialism, including the province of Taiwan and the tributary states of Korea, the Liuqiu Islands and Annam. From this time onward, Chinese nationalism affected in fundamental ways the rendering of geographical space by cartographers in China.

Viewing matters from one angle, we might say that Chinese cartography in late imperial times impeded a more "realistic" understanding of foreign lands and peoples. Certainly it both expressed and reinforced a tributary-based perspective on Sino-foreign relations--one that probably over-estimated the submissiveness and dependency of aliens. Moreover, the emphasis in so many Chinese "world maps" on the great military conquests and unprecedented territorial expansion of the Qing dynasty may well have contributed to an exaggerated sense of self-confidence on the eve of the Western intrusion.

The overwhelming majority of Chinese mappaemundi--including works produced after the Jesuit interlude--depicted "the world" as if the foreigners inhabiting it existed precariously on the fringes of the Chinese empire. Whole continents appeared either as tiny offshore islands or as inconsequential appendages to China's land mass--terrestrial afterthoughts, so to speak. It is not at all clear, however, that a "more "realistic" depiction of foreign lands would have produced a greater sense of military threat. Indeed, some scholars argued that Jesuit-style maps were designed quite deliberately to mislead the Chinese into thinking that the aggressive, avaricious people from "the Great Western Ocean" were farther away than they actually were.

Moreover, we should remember that barbarians were not always marginalized in Chinese "world maps"--even those with certain "traditional" features. Scholars such as Cao Junyi, Yan Yong, Chen Lunjiong, Ma Junliang, and Zhuang Tingfu, for example, made concerted and largely successful efforts to depict foreign territories accurately; and the Manchus, for their own political reasons, produced excellent maps of the Qing empire with Jesuit assistance. In fact, it seems clear that reliable cartographic information existed for those scholars who wanted it, despite Manchu efforts to keep certain types of knowledge to themselves, and notwithstanding the understandable confusion produced by so many different renderings of barbarian space. But the incentive to seek this knowledge out, like the incentive to disseminate it widely, does not seem to have been particularly powerful--at least not until the rise of Western imperialism in the mid-nineteenth century. Quite the contrary, there were political incentives to support the status quo.

This line of analysis assumes a certain pragmatic approach to cartography that obscures other ways of thinking about maps. For many Chinese scholars, maps--world maps in particular--were designed to be appreciated, not simply employed. Although such works had a certain (limited) practical value, they had a much greater emotional appeal. As Liang Zhou put the matter in the introductory remarks to his highly influential map of 1593: "[This work] deals with the grandeur of China's mountains and rivers as well as the excellence of its people, past and present." Emperors and officials may have required certain kinds of finely wrought maps for specific military and administrative purposes, but they also needed large-scale maps as a means taking "spiritual journeys [shenyou] across vast space," in the poetic words of Zhuang Tingfu. Indeed, we know that certain mappaemundi, such as the Liangyi xuanlan (Map for the Profound Observation of Heaven and Earth; 1603) were designed and used expressly for such purposes.

By combining aesthetics, cosmology, history and culture in particularly creative and compelling ways, the makers of Chinese world maps often sought to blur the conventional distinction between actual, lived space and imaginary, idealized space. To put the matter a bit differently, the works they produced played something akin to the role that Geertz ascribes to ritual, linking "the world as lived with the world as imagined."

It is sometimes said that traditional Chinese landscape paintings are not so much depictions of nature as they are statements about the "nature of nature." Similarly, many traditional Chinese mappaemundi are not so much "renderings of the world" as they are cultural statements about the "nature of world." Their purpose, at least in part, was to reinforce certain abiding cultural myths which, in turn, sustained China's self-image--stories about the Central Kingdom's advantageous geographical and cosmological location, its glorious conquests; its impressive explorations; its heros, its, famous landmarks, and its powerful influence on other lands and peoples. Many of the places depicted or referred to in Chinese maps of the world provoked powerful reactions--regardless of whether they were actual locations or purely mythological sites.

This was true not only for points to be found in China Proper, such as the legendary "Sea of Constellations," but also for points beyond. Indeed, one reason for including place names from the Shanhai jing in so many Chinese mappaemundi, was not only to create a sense of comprehensiveness, but also to trigger certain poetic and other literary associations. Every member of the Chinese elite--and probably a number of literate commoners as well--knew of Tao Qian's famous poem, inspired by looking at the illustrations in the Shanhai jing ("Du Shanhai jing"). Two lines suggest its value as a source of imaginary inspiration:

"I view the pictures in the Classic of Mountains and Seas.

A drifting glance [liuguan] encompasses the ends of the universe."

Quite possibly Tao's poem refers to a map (or maps) no longer extant. In any case, dozens of his lines mention specific places in the Shanhai jing--each connected with colorful and well-known images that circulated freely in Chinese folklore. Thus we read:

Far away, the Locust River Range;

At hand, the famed Hanging Gardens Hill.

To the south and west you see the Kunlun Mountains;

The shining atmosphere is hard to match.

The Shanhai jing tells us what every Chinese literatus already knew: that the Locust River Range is topped with the finest gold, silver and jade; that the Hanging Gardens Hill has a tutelary deity with the body of a horse, a human face, tiger stripes and the wings of a bird; and that--as everyone was surely aware--the Kunlun Mountains are the legendary source of the Yellow River, not to mention the abode of the Queen Mother of the West (Xiwang mu). References to even more remote places in Chinese maps--regions inhabited by strange beings such as "The People with Perforated Chests," "The People with One Arm," "The People with Three Bodies," and "The People with One Eye"--naturally conjured up a different set of exotic associations, images shared by elites and common people through media such as encyclopedias and almanacs.

In short, Chinese world maps in late imperial times had several purposes--not all of them either pragmatic or "scientific." Unlike works in the Huangyu quanlan tu tradition, specific claims of territorial jurisdiction in Chinese mappaemundi took a back seat to more general claims of feudal overlordship. And until the twentieth century, mathematical precision was never considered a cartographic end in itself. To be sure, Chinese mapmakers understood the utility and appeal of accurate measurement, and their colleagues in astronomy developed sophisticated instruments that made possible the projections and coordinate systems that Westerners associate with Ptolemaic cartography. But throughout most of the imperial era, they found no compelling reason to conceive of the world as spherical, nor did they see any special merit in drawing all maps "to scale." (After all, the eunuch-admiral Zheng He made his way to the coast of Africa in the early fifteenth century without much difficulty.) Besides, cartographers knew that commentaries could always provide precise geographic details, if they should prove necessary.

Despite a long tradition of sophisticated geographical and cartographic scholarship, an equally long history of foreign exploration (and conquest), and the systematic acquisition of information on barbarians of various kinds, the "outer" world as a whole remained relatively unimportant to the vast majority of Chinese--elites and commoners alike.

In the West, the great voyages of discovery from the late 15th century onward ignited interest in "capturing the world as a single ordered image." But Zheng He's earlier--and in some ways much more impressive--sea voyages had no such effect in China; in fact, they were a source of embarrassment. And whereas the possession and display of a world map or globe from the Renaissance onward in Europe signified that the owner was "a knowledgeable and worldwise citizen," it meant no such thing in imperial China. Thus, until forced to reconsider their craft by new political and cultural priorities, Chinese mapmakers generally made the choice to depict the world not so much in terms of how it "actually" was, but rather in terms of how they wanted it to be.